High speed video action recognition and localization

Inactive Publication Date: 2008-12-18
RGT UNIV OF CALIFORNIA
View PDF2 Cites 57 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Benefits of technology

[0007]Hence, this embodiment provides a system for enabling high-speed real-time video action recognition and localization via use of preexisting motion vectors or other motion field information, which are often byproducts of modern video compression schemes. As discussed more fully below, leveraging precomputed motion vectors and emplo

Problems solved by technology

High-speed action-recognition systems are particularly important in surveillance applica

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • High speed video action recognition and localization
  • High speed video action recognition and localization
  • High speed video action recognition and localization

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0004]Certain embodiments of the invention provide an apparatus for detecting or locating an action in a test video. In a specific embodiment, the apparatus includes a first mechanism for employing motion vector or field information to identify a predetermined characteristic of an action and to provide a first signal in response thereto. A second mechanism includes instructions for determining where in a video sequence a certain motion exists or is likely to exist based on the first signal.

[0005]In a more specific embodiment, the predetermined characteristic includes the motion that is associated with the action. The first mechanism further includes a third mechanism for determining a motion similarity between a first group of pixels in a first frame in a test video and a second group of pixels in a second frame in a query video. The query video includes a representation of the action, and the test video represents the video sequence.

[0006]In this embodiment, the motion similarity i...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

An apparatus for detecting an action in a test video. In an illustrative embodiment, the apparatus includes a first mechanism for receiving a query for a particular action via a query video. A second mechanism employs motion vectors associated with the test video to compute one or more motion-similarity values. The one or more motion-similarity values represent motion similarity between a first group of pixels in a first frame of a query video and a second group of pixels in a second frame of the test video based on the motion vectors. A third mechanism uses the one or more similarity values to search for the particular action or similar action in the test video. In a more specific embodiment, another mechanism aggregates the similarity values over a predetermined number of frames to facilitate estimating where the particular action or version thereof occurs or is likely to occur in the test video.

Description

ACKNOWLEDGMENT OF GOVERNMENT SUPPORT[0001]This invention was made with Government support under Contract Number CCR-0330514, awarded by the National Science Foundation (NSF). The Government may have certain rights to this invention.BACKGROUND OF THE INVENTION[0002]This invention is related in general to video processing and more specifically to action-recognition systems and methods for detecting or locating a particular action in a video. Generally, video action recognition systems compare actions in a query video to many reference videos to detect whether an action in the query video matches the action in a particular reference video. In this manner, a particular action such as, e.g., walking, running, jumping, throwing a punch, or other action performed by an animate or inanimate actor can be identified. Action-recognition systems can be employed in various demanding applications, including video surveillance, collision-avoidance for unmanned aerial vehicles, video editing, video...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/62G09G5/00
CPCG06F17/30781G06K9/00711G08B13/19613G06F16/70G06V20/40
Inventor AHAMMAD, PARVEZYEO, CHUOHAORAMCHANDRAN, KANNANSASTRY, S. SHANKAR
Owner RGT UNIV OF CALIFORNIA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products